In this paper, a novel neural network-based robust control method is presented for a vehicle-oriented problem, in which the main goal is to ensure stable motion of the vehicle under critical circumstances. The proposed method can be divided into two main steps. In the first step, the model matching algorithm is proposed, which can adjust the nonlinear dynamics of the controlled system to a nominal, linear model. The aim of model matching is to eliminate the effects of the nonlinearities and uncertainties of the system to increase the performances of the closed-loop system. The model matching process results in an additional control input, which is computed by a neural network during the operation of the control system. Furthermore, in the second step, a robust is designed, which has double purposes: to handle the fitting error of the neural network and ensure the accurate tracking of the reference signal. The operation and efficiency of the proposed control algorithm are investigated through a complex test scenario, which is performed in the high-fidelity vehicle dynamics simulation software, CarMaker.
The concept of vehicle automation is a promising approach to achieve sustainable transport systems, especially in an urban context. Automation requires the integration of learning-based approaches and methods in control theory. Through the integration, a high amount of information in automation can be incorporated. Thus, a sustainable operation, i.e., energy-efficient and safe motion with automated vehicles, can be achieved. Despite the advantages of integration with learning-based approaches, enhanced vehicle automation poses crucial safety challenges. In this paper, a novel closed-loop matching method for control-oriented purposes in the context of vehicle control systems is presented. The goal of the method is to match the nonlinear vehicle dynamics to the dynamics of a linear system in a predefined structure; thus, a control-oriented model is obtained. The matching is achieved by an additional control input from a neural network, which is designed based on the input–output signals of the nonlinear vehicle system. In this paper, the process of closed-loop matching, i.e., the dataset generation, the training, and the evaluation of the neural network, is proposed. The evaluation process of the neural network through data-driven reachability analysis and statistical performance analysis methods is carried out. The proposed method is applied to achieve the path following functionality, in which the nonlinearities of the lateral vehicle dynamics are handled. The effectiveness of the closed-loop matching and the designed control functionality through high fidelity CarMaker simulations is illustrated.